llama_capi.py previously hand-copied struct field order/types from one
specific llama.cpp commit's llama.h into Python ctypes Structures. A
different llama.cpp build could silently reorder or resize those fields and
corrupt memory rather than raising any error.
Replaced with a cffi "API mode" extension (build_capi.py) that #includes the
user's actual llama.h and links against their actual libllama.so. Struct
layout now comes from real compilation - a genuinely incompatible field
fails the build loudly instead of corrupting memory at runtime. Verified
byte-for-byte identical generation output against the prior ctypes
implementation at a fixed seed, plus the fork/peek logit round-trip check
(0.000000 max diff).
Requires a one-time `python build_capi.py` setup step (needs a C compiler,
which building llama.cpp itself already requires).
llama.cpp-based implementation of the countbayesie Future-Entropy Sampler,
built via a direct ctypes binding to libllama.so. Includes a one-shot CLI
(entropy_cli.py), an OpenAI-compatible server (entropy_server.py), and the
tuning/benchmark scripts used to derive the alpha/confidence-threshold
defaults documented in the README.
Model path and llama.cpp library path are now read from environment
variables (ENTROPY_SAMPLER_MODEL, LLAMA_CPP_LIB) instead of being hardcoded,
and gguf-py is pulled from PyPI instead of a local llama.cpp checkout, so
this runs on any machine with a compatible llama.cpp build and model.
